The gene Ras is mutated to remain in the active oncogenic state in one third of all human cancers, and as high as 90% in pancreatic cancer. In this state active Ras activates the MAPK, PI3K, and RalGEF effector pathways to promote tumorigenesis. However, cancer is a dynamic process, prompting us to test whether Ras must activate the same pathways to initiate tumorigenesis, as it must to maintain tumor growth. We found that while all the effector pathways were needed to initiate tumor growth, once a tumor was established oncogenic Ras signaling was reduced to activation of the PI3K pathway. We propose to identify in aim 1 the proteins downstream of Ras preferentially required to initiate tumor growth, and in aim 2, the proteins required to maintain tumor growth. Our newest preliminary data point toward likely candidates for each of these distinct phases of tumor growth: for tumor initiation- a series of angiogenic factors, and in particular IL6, for tumor maintenance- substrates of AKT, and in particular eNOS. We will focus our initial efforts on these two proteins, addressing whether loss of their function impedes tumor initiation or maintenance in different systems: xenografts of genetically defined human cancer cells given the ease of manipulating such cells, xenografts of human pancreatic cancer cells given the relevance to human cancer and genetically variable backgrounds of the cells, carcinogen-induced skin cancer in mice to address the role of these proteins in a spontaneous Ras-driven model of cancer and lastly, a pancreatic cancer model in mice given the high frequency of Ras mutations in this cancer. Completion of these aims should provide a mechanistic understanding of the Ras signaling required at different stages of tumorigenesis, with the long-term goal of identifying those pathways most crucial to target during the dynamic process of cancer.